Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessReach published sandbox agent-server port (0.0.0.0)
-
DeliverySend unauthenticated HTTP/WebSocket request
-
ExploitInvoke command execution endpoint
-
ExecutionRun arbitrary shell commands as root
-
ImpactRead mounted workspace and plaintext secrets
Vulnerability Assessment
| Weakness |
CWE
ASVS 5.0
V15 - Secure Coding and Architecture Hardware binding mechanism
|
| Exploitation Conditions | Exploitation requires that the deployment use the DockerWorkspace or Apptainer workspace launcher in its default configuration, where the sandbox agent server is bound to all interfaces (0.0.0.0) with an empty session_api_keys list, disabling authentication. The attacker must additionally be able to reach the published container port over the network. Limiting factors: only these two launcher types are affected; deployments that override the default to set session API keys, bind to loopback, or restrict the published port via host firewalling/network segmentation are not exposed. No user interaction or credentials are required against an affected default install. The fix (v1.52.0) publishes Docker ports on loopback rather than all interfaces. |
| Impact | This is a genuine high-priority issue, and the signals are mutually reinforcing rather than conflicting. The supplied CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H, base 9.3) describes network-reachable, low-complexity, unauthenticated exploitation with full confidentiality, integrity and availability impact on the vulnerable system - fully consistent with the description of an insecure-default (CWE-1327, binding to 0.0.0.0 with authentication disabled). The vuln.today decision label is act_now and a vendor patch (v1.52.0) is available, so remediation is urgent and actionable. Tempering the raw score: this is an insecure default rather than a memory-corruption bug, so exploitation is trivial but conditional on the affected workspace launchers being used and the container port being reachable. No EPSS score was provided, the CVE is not listed in CISA KEV, and no confirmed public exploit code is present in the input - so despite act_now priority there is no evidence of active exploitation; the urgency rests on exploitation ease, VulnCheck's advisory, and full technical impact, not on observed in-the-wild activity. |
| Exploit ScenarioDerived | An attacker on the same network as (or with routable access to) a host running OpenHands Software Agent SDK before 1.52.0 with a DockerWorkspace or Apptainer sandbox discovers the published agent-server port, which is bound to 0.0.0.0 with no session API key. They send unauthenticated HTTP/WebSocket requests to run arbitrary shell commands as root inside the sandbox, read the mounted workspace, and exfiltrate plaintext secrets from the secrets store. No public exploit identified at time of analysis, but exploitation is trivial given the missing authentication. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Oct 26, 2026
Unauthenticated remote code execution affects OpenHands Software Agent SDK releases before 1.52.0 when the DockerWorkspace or Apptainer workspace launchers are used in their default configuration: the sandbox agent server binds to 0.0.0.0 with an empty session_api_keys list, so authentication is effectively disabled. A remote attacker who can reach the published container port (CVSS 4.0 base 9.3, assessed as CVSS 3.1 AV:N/AC:L/PR:N/UI:N with high confidentiality, integrity and availability impact) can issue unauthenticated HTTP and WebSocket requests to run arbitrary shell commands as root inside the sandbox, read the mounted workspace, and pull plaintext secrets from the secrets store. Vendor patch v1.52.0 is available and the priority decision is act_now; no public exploit code was identified in the supplied data, though VulnCheck reported the issue.
RemediationAI
Upgrade to OpenHands Software Agent SDK v1.52.0, which publishes Docker workspace ports on loopback instead of all interfaces (PR #5209, commit 7b66584); release notes and artifacts are at https://github.com/OpenHands/software-agent-sdk/releases/tag/v1.52.0. Where immediate upgrade is not possible, remove the exposure the advisory describes: configure a non-empty session_api_keys list so the agent server requires authentication, and/or override the launcher so the sandbox server binds to 127.0.0.1 rather than 0.0.0.0, and stop publishing the container port beyond the host. Because the default publishes the port on all interfaces, network-side restriction of the published container port through host firewalling or segmentation is the remaining compensating control; each of these measures depends on deliberate configuration override rather than the vulnerable default, and any deployment that still exposes an unauthenticated agent server port to untrusted networks remains exploitable. Per the supplied data, v1.52.0 is the vendor-released fix and no public exploit code was identified at time of analysis.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The flaw is a default-permission weakness (CWE-1327: binding to an unintended or unrestricted address) in the OpenHands software-agent-sdk workspace launchers. DockerWorkspace and Apptainer start a sandboxed agent server whose HTTP and WebSocket control interfaces are published on all network interfaces (0.0.0.0) and started with an empty session_api_keys list, which the SDK treats as authentication disabled rather than as a hard failure. Because the agent server's purpose is to accept commands and execute them on behalf of an agent, the command-execution API becomes an unauthenticated remote shell running as root within the sandbox, with visibility into the mounted workspace and the secrets store held in plaintext. The exposure is configuration- and deployment-dependent: only the Docker and Apptainer launcher paths are affected, and only where the published port is reachable from an untrusted network. The upstream fix in v1.52.0 (PR #5209, commit 7b66584) changes Docker port publishing to loopback, so the control interface is no longer exposed off-host by default. The trigger mechanics for individual endpoints are as described in the advisory; no additional memory-safety or protocol-level precondition is indicated by the supplied sources.
Know when Openhands Software Agent Sdk gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Default insecure bind exposes an unauthenticated network endpoint (AV:N/AC:L/PR:N/UI:N) granting root command execution in the sandbox with full C/I/A; impact stays within the sandbox system, so S:U.
Primary rating from Vendor (VulnCheck).
CVSS VectorVendor: VulnCheck
DescriptionCVE.org
OpenHands Software Agent SDK before 1.52.0 contains an unauthenticated remote code execution vulnerability where the DockerWorkspace and Apptainer workspace launchers bind the sandbox agent server to all network interfaces (0.0.0.0) with an empty session_api_keys list, disabling authentication by default. Attackers who can reach the published container port can send unauthenticated HTTP and WebSocket requests to execute arbitrary shell commands as root in the sandbox, access the mounted workspace, and read plaintext secrets from the secrets store.
Executive ViewAI
OpenHands Software Agent SDK, used for sandboxed code execution in AI agent workflows, exposes a critical unauthenticated remote code execution flaw in default DockerWorkspace or Apptainer configurations. Affected deployments let any network attacker run commands as root, access workspace data, and steal plaintext secrets, creating severe risk of data breach and system compromise.
Apply vendor patch.
Threat Intelligence Overview
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.
EUVD-2026-109918
GHSA-rp7h-crm7-h5qj
NVD
CVE.org
EPSS
Exploit-DB